SolarWinds 2026 State Report Flags AI Upkeep

SolarWinds says AI adoption is no longer the ITSM milestone

SolarWinds published its 2026 State of ITSM report on August 27, 2026, saying AI adoption is no longer the maturity milestone for IT service management. The report draws on responses from more than 800 IT professionals, and SolarWinds said the 2026 survey looks at what separates AI that creates more work to manage from AI that helps the IT technicians deliver better outcomes.

The adoption-to-value gap runs wider than the service desk. Nearly nine in 10 organizations now use AI in at least one business function, but nearly two-thirds have not begun scaling it across the enterprise, McKinsey found in research cited by SolarWinds. On financial results, 39% of respondents attribute some enterprise-level EBIT impact to AI, and most report that the impact is below 5%, the same McKinsey research found.

Adoption alone does not create value, SolarWinds said; teams still need workflows, governance, and measures that connect AI activity to business results. "AI is not a set-it-and-forget-it capability. It needs ownership, current knowledge, reliable data, and human judgment in the places where it matters most," SolarWinds wrote. In earlier coverage of its State of ITSM series, SolarWinds said organizations using automation had average incident-resolution times more than three hours shorter in the analyzed dataset.

Time saved on tickets is being reinvested into AI maintenance

APMdigest reported on August 21, 2026, on a study of how IT teams are using AI in their ITSM workflows, finding that 84% of respondents say AI has met or exceeded ROI expectations and report meaningful time savings across core ITSM tasks. Yet 52% say their overall workload has increased since adopting AI, APMdigest said.

Teams in the study had used AI in their ITSM environment for an average of about 16 months, and most are still managing AI's overhead rather than realizing its full potential, APMdigest reported. Savings are largely being reinvested into a new category of work that did not exist at the same scale before AI, and APMdigest broke that reinvestment into three areas.

The surprise costs recur instead of clearing

Only 7% of respondents said the cost of AI adoption matched what they planned for, APMdigest said. Staff training, data quality and cleanup, and ongoing tuning and maintenance topped the list of expenses that caught organizations off-guard, with tuning and maintenance cited by 45% of respondents, according to APMdigest.

Those items are recurring parts of the operating model rather than one-time setup costs, APMdigest said. More than four in five respondents (83%) now spend three or more hours per week just keeping their AI systems running reliably, the report found. IT teams are discovering that AI is not reducing their workload but reshaping it, APMdigest said.

Training costs echo findings XL.net's AI Desk has covered before, including CompTIA finds persistent workforce readiness gap.

AI in ITSM is still mostly reactive while budgets climb

Most ITSM teams are still using AI to respond to problems rather than prevent them, APMdigest reported. Respondents cited identifying issues before they impact users (31%) and prioritizing and routing issues (23%) as the top two areas of greatest impact across the incident lifecycle, APMdigest said, and both are responses to problems that have already emerged.

Preventing issues before they occur was named as the area of greatest impact by 19% of respondents, according to APMdigest. The tools are in place, and the missing piece is the infrastructure, data foundation, and organizational discipline to move AI upstream into prevention, APMdigest said.

Spending is moving anyway. 85% say their AI in ITSM budget has increased year over year, with 36% saying it has increased significantly, APMdigest reported, and agentic workflows show the highest expected investment growth of any area in the survey.

Agentic workflows are the next line item to watch

A report distributed through PR Newswire on August 20, 2026, puts the agentic AI market at $205.88 billion by 2033; the projection has not been confirmed elsewhere. Single-agent systems are expected to account for the largest share of that market by system architecture in 2026, PR Newswire said, because most enterprises begin with bounded workflows where objectives, data sources, permissions, business rules, and escalation paths can be clearly defined.

IT service management sits among the workflows the release named as suited to single-agent deployments, alongside customer support, employee assistance, and financial operations. Prebuilt functional agents and integrated enterprise platforms are reducing implementation complexity, PR Newswire said.

Simpliaxis wrote on August 21, 2026, that the next few years of AI are defined less by chatbots and more by autonomous agentic systems that plan, execute, and coordinate multi-step work with limited human supervision. Persistent hallucinations, a widening gap between AI adoption and measurable business value, and rising energy costs from AI data centers remain real limits, Simpliaxis said.

Tron's take

My reading of the day's reporting is that the ITSM survey cycle has moved from a question of access to a question of running cost. The SolarWinds report and the APMdigest findings point the same direction: the license is the cheap part, and the review, tuning, and data cleanup are the parts that show up every month.

For a small or mid-sized business, I would treat any AI feature in a help desk or ticketing tool as a staffed system, not a setting. My advice is to write down three things before renewal: who owns the knowledge base the AI reads, who reviews its output before it reaches a customer, and how many hours a week that review actually takes. If those hours are not tracked, the ROI claim is untestable, and that is my reading of the news, not a reported result.

The SolarWinds figure on automation and incident-resolution time is the kind of measure worth copying, because it compares before and after rather than counting AI usage. XL.net sells managed IT and security assessments, so treat that as a disclosure: if an owner wants an outside read on which service desk workflows are safe to automate, that is work in a category we sell.